Naval Research LaboratoryWashington, DC 20375-5000

AD-A241 241 NRL Memorandum Report 6896

Post-Flashover Fires in Simulated ShipboardCompartments: Phase II-Cooling of Fire

Compartment Boundaries

J. T. LEONARD, C. R. FULPER, R. L. DARWIN.* G. G. BACK.**R. J. OUELLETTE,** AND J. L. SCHEFFEY,** R. L. WILLARD**

Navy Technology Center for Safety and SurvivabilityChemistryv Division

*Naval Sea Systems CommandWashington, DC

**Hughes Associated, Inc.,

Wheaton, Maryland

September 19, 1991 D 11C- ELEcT,,a 199L

91-12262

Approved for V-"blic releac C: ibution unlimited.

Form Approved

REPORT DOCUMENTATION PAGE I oem No d74,08

*,.,t'. re1onai b.den for thr t h tndia n Of fo0t.01 o4 1 eSmll-ated Io iie, age I hour oer respon,, incIuding the time fot retv.CW.fl ih$w unions. w(c(hnq e..II,4 & a c"f wosrcga~h en . 1 -6-la " the d41ta .ftded.. :-Oleti¢ &nd ,ev* ,nq 1h (Oth toln of ,onsnat O. Send cornlnts rea thn bu rden b t%,soneate or any Olher as's( of IhicOnedofIn Of inlOfnatO. .wluding su ettois ior virilusng tIhm buidn to Wasthnglon "e&adqcuarle l rs. OattOrwte I o Inlormt.on Operations .nd Repons. 12 15 Jete esonO4hh, hsighway. Slute 1204. Ahington. VA 222024302. and to the Offite of Manq4cne.1 and Oidget. Pae f'rOrk RId-.OlO PrOl (l (0704-018). W.ast.*ntOn. OC 20 3

i. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

1991 September 19 Interim4. TITLE AND SUBTITLE S. FUNDING NUMBERS

Post-Flashover Fires in Simulated Shipboard Compartments:PHASE II - Cooling of Fire Compartment Boundaries

6. AUTHOR(S) PE - 63514N

J. T. Leonard, C. R. Fulper, R. L. Darwin.* G. G. Back.** S1565-SL

R. J. Quellette.** J. L. Scheffey.** R. L. Willard** J.O. # - 61-2961-0-1

7. PERFORMING ORGANIZATION NAME(S AND ACD-ESS(ES) 8. PERFORMING ORGANIZATION

REPORT NUMBER

Naval Research Laboratory NRL MemorandumWashington, DC 20375-5000 Report 6896

9. SPONSORING/MONITORING AGENCY NAME(S) AND AODRESS(ES) 10. SPONSORING/MONITORINGAGENCY REPORT NUMBER

Naval Sea Systems CommandWashington, DC 20362-5101

11. SUPPLEMENTARY NOTES

*Naval Sea Systems Command, Washington, DC 20375-5000.**Hughes Associated. Inc., Wheaton, Maryland 20902

12a. DISTRIBUTION/AVALABWLITY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release; distribution unlimited.

13. ABSTRACT (Maximum 200 words)

A series of tire tests was conducted in simulated shipboard compartments to evaluate the effectiveness ofvarious compartment cooling techniques in preventing the spread of fire both horizontally and vertically. The

compartments were steel cubes, 2.4 x 2.4 x 2.4 m (8 x 8 x 8 ft), arranged such that the fire compartment was in

the center with two cubes on either side and one compartment was directly overhead. All compartments wereinstrumented with thermocouples and heat flux transducers. The fire threat was a 2 gpm JP-5 fuel spray fire tosimulate instant flashover of the fire compartment. Both manual and installed water spray nozzles were evaluatedfor cooling efficiency.

An application rate of 2.04 pm (0.05 gpm/* ) was found to be the minimum required to cool both horizontaland vertical compartment boundaries to 100C (212F). Although cooling of horizontal boundaries was notaffected by application technique, cooling of vertical boundaries was highly dependent on technique. Maximumcooling of vertical boundaries was achieved by applying water tangentially to the boundary surface in sheets or

large droplets with nozzles that produce a fan shaped water spray pattern. Maximum cooling was achieved when

the nozzle was oriented to provide complete coverage of the bulkhead by the water spray pattern, and the water

droplets remained on the heated surface for a sufficient time to absorb the heat. Full cone nozzles spraying

perpendicular to the heated surface provided some cooling for medium size droplets at higher application rates,while fine atomizing nozzles did very little to reduce the bulkhea-d surface temperature.

14. SUBJECT TERMS 15. NUMBER OF PAGES44

Compartment fires, Compartment cooling 16. PRICE CODE

Post-flashover. Fire boundaries

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACT

UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED ULNSN 7540-01-280-5500 Standaid Form 298 (Rev 2-89)

b, &"%I %to 119.16

CONTENTS

Page

1.0 BACKGROUND ..................................... 1

2.0 INTRODUCTION ..................................... 1

3.0 OBJECTIVES ...................................... 2

4.0 APPROACH ........................................ 2

5.0 GENERAL TEST DESCRIPTION ......................... 5

5.1 M ock-Up ..................................... 55.2 Fuel System .................................. 55.3 Instr-jm entation ................................ 10

5.3.1 Air Thermocouples ........................ 105.3.2 Wall Thermocouples ....................... 105.3.3 Heat Flux Transducers ..................... 105.3.4 Loa Cells ............................... 105.3.5 Com puter ............................... 105.3.6 Video and 35 mm Still Cameras ............... 13

6.0 PROCEDURES ..................................... 13

6.1 General Procedures ............................. 136.2 Prelim inary Tests ............................... 13

6.2.1 Variable Flow Test ......................... 136.2.2 Total Periphery Cooling Test ................. 146.2.3 Water Curtain Test .......................... 16

6.3 Horizontal Boundary Cooling Tests ................. 166.4 Vertical Boundary Cooling Tests ................... 176.5 Manual Cooling Techniques ....................... 17

7.0 RESULTS AND DISCUSSION ........................... 19

7.1 Prelim inary Tests ............................... 197.1.1 Variable Flow Test ......................... 197.1.2 Total Periphery Cooling Test .................. 197.1.3 W ater Curtain Test ......................... 19

7.2 Horizontal Boundary Cooling Tests .................. 237.j Vertical Boundary Cooling Tests ..................... 237.4 Manual Cooling Techniques ....................... 29

iii

CONTENTS (Continued)

PAWe

8.0 SUMMARY AND CONCLUSIONS ........................ 34

9.0 RECOMMENDATIONS ................................ 35

10.0 REFERENCES ..................................... 36

APPENDIX A - Tpt nt . . . .

APPENDIX B - Comments to Naval Ships Technical ManualChapter 555, Section 5 ................ B-1

-eces0 ion ro

NTIS GFA&IDTIC TAL

unannounced 0

stiricatiO

... iailitl Codesfai nd/or

Dist Speci-

iv

FIGURES

Fig. Pae

1. Fire compartment air temperatures (design fire) ............. 3

2a. Upper compartment average deck temperatures ............. 4

2b. Adjacent compartment average bulkhead temperatures ........ 4

3. Ship compartment mock-up (elevation view) ............... 6

4. Ship compartment mock-up (plan view) ...................

5. Fueling system ..................................... 8

6. Nozzle assembly .................................... 9

7. Instrumentation layout ................................ 11

8. Instrum entation detail ................................ 12

9. Boundary cooling nozzle assembly ...................... 15

10. Water spray nozzle orientations for vertical boundary cooling tests:(a) center sideways, (b) deck up, (c) ceiling down, and(d) perpendicular to heated surface (conical spray pattern ...... 18

11 a. Variable flow test-horizontal deck average surface temperatures . 20

11b. Variable flow test-vertical bulkhead average surface temperatures 20

12. Fire compartment air temperatures during aggressive cooling ofall boundaries ...................................... 21

13. Water curtain test - radiant heat flux ...................... 22

14a. Fire compartment east wall average temperatures at low applicationrate (3 gpm ) ........................................ 28

14b. Fire compartment east wall average temperatures at high applicationrate (60 gpm ) ....................................... 28

15. Vaporization rate as a function of plate temperature .......... 30

mn

FIGURES (Continued)

Figq. Paaie

16a. Fire compartment east wall average temperatures when coolingwith 30 gpm vari-nozzle (Test 121) ....................... 33

16b. Fire compartment east wall average temperatures when coolingwith 95 gpm vari-nozzle (Test 122) ....................... 33

vi

TABLES

Table Pae

1 Horizontal Boundary Cooling Tests ...................... 24

2 Vertical Boundary Cooling Tests ........................ 35

3 Manual Cooling Techniques ............................ 31

vii

POST-FLASHOVER FIRES IN SIMULATED SHIPBOARDCOMPARTMENTS: PHASE H - COOLING OF FIRE

COMPARTMENT BOUNDARIES

1.0 BACKGROUND

The Internal Ship Conflagration Control (ISCC) program was initiated to addressissues raised by the missile-induced fire on the USS STARK. The overall objectives ofthe program are to develop guidance to the Fleet on the control of horizontal andvertical fire spread and to develop concepts and criteria for new ship design. Theprogram includes small scale fire tests in a simulated shipboard compartment andlarge scale tests in full scale compartments aboard the Navy's fire test ship, the ex-USS SHADWELL, in Mobile, AL.

As part of the ISCC program, an analysis was conducted to characterize theconditions occurring during post-flashover compartment fires in a simulated shipboardcompartment [1]. In this test series, the thermal conditions occuring in the firecompartment were quantified, and the likelihood and estimates of fire spread rateswere determined. In the present study, optimum water cooling quantities (flow rates)for containing post-flashover compartment fires were determined, and manual coolingtechniques were developed for preventing vertical and horizontal fire spread.

2.0 INTRODUCTION

According to the current shipboard firefighting procedures, the first crewman toarrive on the scene must decide whether to attack the fire or begin containmentprocedures. If the fire is still small, then an attack on the fire would be appropriate. If,however, the fire is large, as in the case of a missile-induced fire or a fire fed by apressurized fuel source, then initial actions should be taken towards isolation andcontainment of the fire. After the fire has been contained and additional personnelbecome available, firefighting and containment can then proceed together.

Containment procedures require that boundaries be established around theperiphery of the fire including both above and below the fire compartment. It is veryimportant that these boundaries be established as quickly as possible. An analysisconducted in the post-flashover characterization test series [1] showed that fire mayspread both horizontally and vertically within five minutes of flashover. Since fire tendsto spread faster vertically than horizontally, emphasis should also be placed onestablishing the boundary above the fire compartment. Typical procedures employedto establish these boundaries include cooling of bulkheads and decks, cooling ofordnance, aind wetting of combut:1,b',c. to prevent ignition.

Manucrpt approved June 27. 1991.

Chapter 555, Section V, of the Naval Ships Technical Manual on fire fighting [2]suggests that when cooling bulkheads or extinguishing interior fires, short waterbursts are preferred over flowing water continuously. The practice cf using continualwater flow produces large amounts of steam, reduces visibility, and creates potentialflooding and stability problems. If this excess water is not removed from the ship byeither cutting holes in decks and bulkheads or by the use of portable dewateringequipment, the ship may develop a list. In many cases, depending on the sea state, asevere list can be more threatening to the ship than the fire itself.

This water efficiency test series served as an initial investigation into optimumwater cooling quantities (flow rates) and techniques for preventing vertical andhorizontal fire spread. These initial quantities and techniques will be further refined infull scale fire tests on the Navy's Fire Research and Test Ship, the ex-SHADWELL.

3.0 OBJECTIVES

The objectives of this test series were to develop techniques and proceduresso that water may be used efficiently to cool bulkheads and decks to establish fireboundaries in order to prevent vertical and horizontal fire spread, while minimizing thehazards of flooding.

4.0 APPROACH

T, iesc . , t~zts r c-,ducted ,th. intermediate scale shipboardcompartment mock-up developed for the post-flashover characterization test series[1]. The tests were conducted at the Naval Research Laboratory's (NRL) ChesapeakeBay Detachment (CBD) fire test facility. The mock-up was designed to withstandmutlple high intensity fires while still approximating the thermal characteristics (i.e.,heat transfer characteristics) of a typical shipboard compartment. Aoolication ratesand techniques were evaluated against the "design fire" used in the post-flashovercharacterization test series. The "design fire" was developed to approximate thetherma; insult resulting from a nearly instantaneous, post-flashover compartment fire.Such a fire might result from a missile strike in which the warhead fails to detonate,but burning missile propellant is strewn about the compartment. Air temperatures ofover 100(C (1 8320F) were measured in the fire compartment during the design fire asshown in Fig. 1. Bulkhead and deck surface temperatures were observed toapproach 8000C (1 4720F) as shown in Figs. 2a and 2b. By using the design fire asthe test fire in this evaluation, the application rates and boundary cooling techniqueswere evaluated against worst case bulkhead/deck thermal conditions.

2

LbJ0

C)) 0.

00

ULJ

V)V

LL E

E -

LLJ~ C%

z ILLli 0

0 0 0 0 0 0 0 0 0 0 0 000 0- 0 0 0 0D Lt 0* 0) 0

w - 0 " L .- n C

(0) ainIDJadwai

1200

1100 - Exposed

1000 -- Unexposed

900iTe 800mp 700

ur[ 00 510 5420 5.30 35.40 45.5

e

Tim (inues

1200

r1400-Exoe

300

e 200

100

0

0 5 10 15 20 25 30 35 40 45 50

Time (minutes)

Fig. 2a - Upper compartment average deck temperatures

12001100 - Exposed

1000 Uneposed

Te 800mp 700

r 600

a 0 ............ .,.

r 400-e 30

PC 200-100

0 5 10 15 20 25 30 35 40 45 50

Time (minutes)

Fig. 2b - Adjacent compartment average bulkhead temperatures

4

The systematic approach incorporated in the development of optimum water

cooling techniques was as follows:

A series of preliminary tests designed to answer specific questions

pertaining to boundary cooling were conducted to serve as a baseline inthe latter test series.

* The minimum application rate for cooling the ideal case (flat, level,

horizontal decks) was determined.

Incorporating the above application rate, the spray configurations

required to produce maximum cooling of vertical bulkheads withminimum water flow were developed.

* Using the above knowledge, the techniques for cooling decks and

bulkheads using standard Navy hardware were then determined.

5.0 GENERAL TEST DESCRIPTION

5.1 Mock-Up

The full size mock-up constructed at CBD for the post-flashover characterizationtest series [1] was used in this evaluation. The mock-up consisted of four 2.4 x 2.4 x2.4 m (8 x 8 x 8 ft) cubical enclosures, three cubes long and two cubes high in thecenter as shown in Figs. 3 and 4. The mock-up was constructed of 0.95 cm (3/8 in.)thick steel plates. Stiffeners having 'T' shape cross sections were welded vertically tothe center of each wall in all compartments. The outside lower compartments eachcontained two 66 x 167.6 cm (26 x 66 in.) openings, one to the outside air and one tothe center compartment. The upper compartment contained one door opening to theoutside air and one (0.61 m (2 ft) diameter) hatch opening in the overhead. Thecenter compartment contained four doors, one to each of the adjacent compartmentsand two to the outside air.

5.2 Fuel System

The fueling system and nozzle assembly are shown in Figs. 5 and 6. Thissystem was used to achieve post-flashover fire conditions in the fire compartment asquickly as possible. The fueling control station was located 6.1 m (20 ft) behind thefire compartment. Ouick operating quarter-turn valves were installed for manualshutdown of the system. A nitrogen system was installed to pressurize the fuelstorage tank and to flush out the fuel system after each test. The fueling station wasmanned at all times during testing.

5

hT.-

CO

X- x

CJ6C'J 0

I NQ

6 0

E~~cc

V, o,

CCv,

C6 'I6

ci cv-

EAI ..

CDC) 2: a) C)

TE

0

a) ca)0L

CL

E--

HEEoC-.

45.4 kg- (100 lb.)Nitrogen cylinder

Flexible hose' ' -L3_1/4 Turn ball valves

"-Load cell assembly

1891 114ga1

(570ga 3 gal.)Fuel Fueltank tanktank

1 1/4 Turn ball valvesStrainer

Flexible_____hose 1.9 cm

Pe r (3/4 in.) GlobePressure___________ valvegage

Compartment wall -. 9 cm

~Pan

\,,\1.2 x 1.2 x 0. 1 m'" (4 x 4 x .3 3 f t)

Fuel nozzle

Fig. 5 - Fueling system

8

T" 14 5 / I , "".,

Pan surface

Bete FF125145 nozzle Bete FF125145 nozzleside elevation front elevation

Nozzle spray

Nozzle location

Fuel pan - schematic

Fig. 6 - Nozzle assembly

9

5.3 Instrumentation

The same instrumentation scheme installed for the post-flashovercharacterization test series was incorporated in this evaluation (Figs. 7 and 8).

5.3.1 Air Thermocouples

Thermocouple trees installed in all con ilrtments provided air temperaturemeasurements. All thermocouples used in the test series were Type K. Inconel-sheathed thermocouples were installed in the fire compartment, while hightemperature glass braided thermocouples were installed in the adjacentcompartments.

5.3.2 Wall Thermocouples

Matrices of thermocouples installed on both exposed and unexposed surfacesof the bulkheads and decks bounding the fire compartment provided information onthe energy conducted through the steel plates and the quantities of heat beingremoved from the plate during the cooling process. Inconel-sheathed thermocoupleswere used to measure surface temperature. These thermocouples were fastened tothe boundaries by drilling a small hole and peening the end of the thermocouple tothe surface.

5.3.3 Heat Flux Transducers

Radiation and total heat flux data collected from each compartment served asan indicator of the energy being removed from the boundaries during the coolingprocess. High range (330 kW/m2 (30 BTU/ft2 s)) transducers were installed in theceiling of the fire compartment and medium range (110 kW/m 2 (10 BTU/ft2 s))transducers were installed in each of the adjacent compartments.

5.3.4 Load Cells

Load cell assemblies installed under the fuel storage tank provided mass lossrates from which fuel flow was calculated.

5.3.5 Computer

An IBM compatible computer, a data acquisition system produced by MetrabyteCorporation consisting of one DAS-8 and seven EXP-16 interface cards were used toscan the above inst'uments in ten second intervals. A commercial software package(Lab Tech Notebook) was used to drive the entire system. The data were stored onfloppy disks (ASCII format) to be analyzed and manipulated at a later date.

10

Upper Compartment

OCRAD in ceiling

ATC

OTC

oRADI * ATCOTHIn ceiling *

THFa ATC BTB- ATC RA-

BTC" ATC RRTHF

ATA

West Compartment Fire Compartment East Compartment

ATC - Air Thermocouples (Tyee K)DTC- Deck Thermocouples (Type K)BTC - Bulkhead Thermocouples (Type K)GS -Gas Sampling (02,CO C02)RAD - Radiometers 110 kW/mrP2 (10 BTU/ft 2 sec) RangeTHF - Total Heat Flux Transducers

110 kW/m 2 (10 BTU/ft 2 sec) Range'330 kW/m 2 (30 BTU/ft 2 sec) Range

Fig. 7 - Instrumentation layout (plan view)

11

Instrumentation Detail

2.13 m1.07 m (7.0 ft)

0.30 M .5 ft -- 1.52 m(1.0 ft) Jj (5.0 ft)

<* WTC

CTC 2.13 m(7.1 ft) RAD THF

0.' (7. m XmT T

O f91m 6 cm (2 ft) 1.2m3. t) * _L (4 ft)

30 cm (1 ft)

Ceiling/Floor Typical WalThermocouples Instrumentation

(plan) (elevation)

ATC

_I30 cm (I ft) Typ.

T

15 cm (6 in)

TAir Thermocouples

(elevation)

Fig. 8 - Instrumentation detail

12

5.3.6 Video and 35 mm Still Cameras

Visual recordings, both still and motion, were made of each test. Theserecords serve as q means of estimating the level and volume of steam produced bythe water application and were archived to serve as a visual record.

6.0 PROCEDURES

6.1 General Procedures

Upon completion of the pre-test checks of instrumentation, fueling system, andsafety equipment, the area was cleared for the start of the test. Once allcrewmembers were in position, the data acquisition system, video cameras, andstopwatches were all started marking the beginning of the test. These systems wereactivated one minute before ignition to collect background data and to record theignition information. Thirty seconds after activation of the above systems, a smalltorch was !i and placed in the fuel pan in the fire compartment At ,ne minute intothe test, the fuel system was charged and the fuel flow rate was adjusted to thedesired amount (7.5 Ipm (2 gpm)). The cooling or water application system was notactivated until eleven minutes into the test. At the eleven minute mark, steady stateconditions were approached; i.e., the fire compartment temperatures reached 1 0000C(18320F) and bulkhead temperatures, exposed (fire side) and unexposed surfacetemperatures, reached 6000C (111 2°F) and 5000C (9320F) respectively. The coolingsystem was then activated for the desired amount of time (usually three minutes).Once the required information was collected, the fuel and cooling systems wereshutdown and the test was terminated.

6.2 Preliminary Tests

A series of tests were conducted to: aid in the refinement of test procedures,check instrumentation, and to answer specific questions about water application,steam production, and fire spread prevention. These tests included the following:

6.2.1 Variable Flow Test

The objective of this test series was to determine the minimum application ratefor effective cooling of both vertical and horizontal boundaries.

Theoretically, the quantity of water required to cool a compartment boundary isrelatively low in comparison to flow rates of vari-nozzles, all purpose nozzles, andapplicators. These theoretical quantities of water are based on the conversion ofwater to steam independent of how the water is applied, the type c; ",,e c, t-surface orientation. Estimates developed by NIST [3] show that an aluminum deckabove a fire compartment having 50 kw/m2 (4.5 BTU/ft2 sec) of heat being conductedthrough it can be cooled to below 100°C (212"F) with an application rate of 1.34Ipm/m 2 (0.033 gpm/ft2). This assumes the water is applied in an infinitely thin layer

13

evenly across the deck and at such a rate that all the water is turned into steam. Thiscorresponds to an application efficiency of 100%. MPR [4] suggests that a higherapplication rate of 2.44 Ipm/m 2 (0.06 gpm/ft) is a better approximation. Using a safetyfactor of 0.66 to produce an application rate of 4.07 Ipm (0.1 gpm/ft), this means thata "typical" fire fighter using a standard Navy 360 Ipm (95 gpm) vari-nozzle couldeffectively cool a 93 m2 (1000 ft2) boundary.

The unknown in the above theoretical calculations is the efficiency with whichthe water is applied to the surface (application efficiency). The assumption of 100%efficiency made by both NBS and MPR may be realized or nearly realized whencooling a flat, unobstructed, horizontal deck. The justification ior this assumption isthat for any application rate greater than the theoretical value, a layer of water willdevelop covering the entire deck independent of application technique. Applicationefficiency does not come into play until attempting to cool a vertical bulkhead, or adeck which is no longer level due to ship listing or rough seas.

Apip.~on rates from 0.4 Ipm/m 2 (0.01 gpm/ft) to 4.0 Ipm/m2 (0.1 gpm/ft2) wereevaluated in this test series to determine the critical application rates required toeffectively cool both horizontal and vertical boundaries. Application rates weredetermined by dividing the flow rate of the nozzle by the heated surface area. Waterwas applied to the heated surface using the fixed system shown in Fig. 9. The systemconsisted of one spray nozzle (Model TF14FC manufactured by Bete Fog Nozzle, Inc.)which produced a solid cone water spray pattern. The TF174FC has a K factor of 1.4as defined in the following equation,

flow (gpm) = K Vpressure (psi)

The nozzle was installed 1.2 m (4 ft) from the heated surface to provide completecoverage by the water spray.

The system was activated eleven minutes into the test as described above. Thesystem was initially set for an application rate of 0.4 Ipm/m 2 (0.01 gpm/ft2). After thefirst minute of cooling, the application rate was increased by 0.4 Ipm/m 2 (0.01 gpm/ft2)each minute thereafter until an application rate of 4.0 Ipm/m 2 (0.1 gpm/ft) was reachedand the test was terminated.

6.2.2 Total Periphery Cooling Test

The objective of this test series was to determine whether fire compartmenttemperatures could be reduced to any extent by aggressively cooling all sides of thecompartment. Five standard Navy 360 Ipm (95 gpm) vari-nozzles were used to coolthe boundaries in this evaluation. A portable hydrant was used to supply water for thefive hand lines used in this test. The portable hydrant was connected to a domestichydrant via 6.3 cm (2.5 in.) fire hose. The design fire was allowed to burn for 15minutes with agressive cooling occuring the last 5 minutes.

14

/7I//

/ /

/ / /

3.8 x 1.9 cm / ,/ / /

(1.5 x .75 in) Steel plate --

bushing30.5 x 30.5 x .95 cm(12 x 12 x .375 in)

1/4 turn valve \\1 '"\\ \ /

To water \ ',

supply V

NFT Fire hose connection 1.9 cm (.75 in)3.8 cm

(1.5 in)

Vertical boundary cooling nozzle assembly

Horizontal boundary cooling nozzle assembly

Fig. 9 - Boundary cooling nozzle assembly

15

6.2.3 Water Curtain Test

The objective of this test series was to determine if low flow water curtainscould block heat radiating from hot bulkheads as suggested in Reference 5.

The nozzle assembly developed for the variable flow test series was used in thisevaluation. A nozzle that produced a uniform wall of water was selected for thisanalysis (Bete Model TF8XW, K = 0.5). The nozzle was positioned to produce a watercurtain 0.92 m (3.0 ft) in front of the heated bulkhead. A radiometer was positionedbehind the nozzle assembly 1.2 m (4.0 ft) from the bulkhead. Water from the nozzledid not come into contact with the heated surface. The procedures were the same asabove; i.e., the nozzle was charged eleven minutes into the test and allowed to flowfor 5 minutes. The nozzle was flowing 11.4 Ipm (3.0 gpm) at a pressure of 0.35 MPa(50 psi).

6.3 Horizontal Boundary Cooling Tests

The objective of this test series was to determine how water should be appliedto a flat, horizontal deck to achieve maximum cooling with minimum application rates.

The nozzle assembly developed for the variable flow tests was used in thisevaluation. Nozzles having various water spray patterns were purchased from BeteFog Nozzle, Inc. The following nozzles and corresponding spray patterns were usedin this evaluation:

Nozzle K factor Spray Pattern

FF93145 0.24 Flat fan-shapedFF125145 0.39 Flat fan-shapedTF10FC 0.65 Full cone, medium drop sizeTF8XW 0.41 Flat 3600 wallTF14FC 1.25 Full cone, large drop sizeP120 0.38 Full cone fine atomization

The flow characteristics of the nozzles were selected based on the criticalapplication rate of 2.04 Ipm/m 2 (0.05 gpm/ft2) over a bulkhead/deck surface area of 6m2 (64 ft). These nozzles were selected to produce this application rate at or belowthe domestic water main pressure of 0.35-0.41 MPa (50-60 psi). In most cases, theflows were low enough to ensure that the water main and the nozzle pressures wererelatively equal. Only during the tests with higher application rates were the nozzlepressures different. The nozzles were evaluated over a range of flow rates and nozzleorientations to achieve maximum cooling with minimum water. The nozzle was alwaysoriented in such a way to assure total coverage of the heated surface by the nozzle'swater spray pattern. A baseline comparison of two vari-nozzles (Akron Brass Co. 360pm (95 gpm) Model No. 3019 and Akron Brass Co. 114 Ipm (30 gpm) Model No.4508) was also conducted.

16

6.4 Vertical Boundary Cooling Tests

The objective of this test was to determine how water should be applied to avertical boundary to achieve maximum cooling with minimum application rates. Thenozzles, nozzle assembly, and procedures developed for the horizontal test serieswere incorporated in this evaluation. The nozzles were evaluated in the followingorientations:

Center Sideways - The nozzle was installed on the left side, one half way

up the wall spraying across (to the right) of the horizontal centerline asshown in Fig. 10(a).

Deck-Up - The nozzle was installed in the middle of the bulkhead at the

deck leve; spraying up the vertical centerline as shown in Fig. 10(b).

Ceiling-Down - The nozzle was installed in the middle of the bulkhead at

the ceiling level spraying down the vertical centerline as shown in Fig.10(c).

The nozzles that produced conical water spray patterns were installed 1.2 m (4ft) from the center of the boundary, spraying perpendicular to the heated surface asshown in Fig. 10(d).

6.5 Manual Cooling Techniques

The objective of the manual cooling test series was to determine how to coolbulkheads and decks using standard Navy 360 Ipm (95 gpm) and 114 Ipm (30 gpm)vari-nozzles. The analysis combined continuous water flow, pulses of water, variationsin water spray patterns, and sweeping maneuvers to determine an optimum coolingtechnique. Other lower flow nozzles were also evaluated in this test series.

The procedures for the manual cooling tests were similar to those usedpreviously. Upon completion of the pre-test checks of instrumentation, fueling system,and safety equipm.;nt, the area was cleared for the start of the test. Once allcrewmembers were at their stations and the fire fighters were dressed and in position,the data acquisition system, video cameras, and stopwatches were all started markingthe beginning of the test. These systems were activated one minute before ignition tocollect background data and to record information pertaining to the ignition sequence.Thirty seconds later, a small torch was lit and placed in the fuel pan in the firecompartment. At one minute into the test, the fuel system was charged, and the fuelflow rate was adjusted to the desired amount (7.5 Ipm (2 gpm)). Manual coolingtechniques were initiated eleven minutes into the test as in the previous tests. The firefighters, dressed in turnout gear, began cooling of the fire compartment bulkheadthrough the adjacent compartment east of the fire compartment. During the majorityof the tests, the fire fighters executed their cooling techniques from the base of thedoor leading into the east adjacent compartment. Once the required information wascollected, the fuel system was shutdown and the test was terminated.

17

(ci) (b)

(c) (d)

Fig. 10 - Water spray nozzle orientations for vertical boundary cooling tests:(a) center sideways, (b) deck up, (c) ceiling down, and

(d) perpendicular to heated surface (conical spray pattern)

18

7.0 RESULTS AND DISCUSSION

7.1 Preliminary Tests

7.1.1 Variable Flow Test

The results from the variable flow tests conducted on the horizontal deck abovethe fire compartment are shown in Figs. 11 a and 11 b. As shown in Fig. 11 a, 2.04Ipm/m 2 (0.05 gpm/ft) was determined to be adequate to lower the unexposedhorizontal surface temperatures of the upper deck to 100'C (212'F). As predicted,application rates above 2.04 Ipm/m2 (0.05 gpm/ft2) resulted in a layer of residual waterdeveloping on the deck providing effective cooling independent of water applicationtechnique. As the depth of the water layer increased, the exposed surfacetemperature of the deck was observed to decrease.

The results of the vertical bulkhead variable flow tests are also shown in Fig.11 b. As shown in this figure, vertical boundary cooling is strongly dependent onapplication technique. Although the technique used in this test series never cooledthe bulkhead below 1500C (3020F), the relation between application rate and bulkheadsurface temperature can still be determined. The cooling gained through increasedapplication rate becomes minimal for application rates above 2.04 Ipm/m 2 (0.05gpm/ft). This application rate was also the rate determined in horizontal deck variableflow test.

7.1.2 Total Periphery Cooling Test

The temperatures recorded in the fire compartment, while cooling all fourboundaries, are shown in Fig. 12. As shown in Fig. 12, aggressively cooling of allboundaries had no effect on fire compartment temperature. The volume of flameproduced by the design fire is adequate to fill the entire compartment producingextremely high temperatures independent of losses through enclosure boundaries.Less severe fires may have produced different results.

7.1.3 Water Curtain Test

The radiant heat measured during the water curtain test is shown in Fig. 13. Asshown in this figure, the radiant exposure 1.2 m (4.0 ft) from the bulkhead wasreduced from 17.0 kW/m 2 (1.6 BTU/ft2 sec) to 3 kW/m 2 (0.28 BTU/ft2 sec) during thistest. Although it was determined that low flow water curtains substantially reduceradiant heat exposures, it is doubtful that this technique could be applied to shipboardfire situations due to obstructions and overall clutter in most compartments. This 75%reduction in radiant heat flux exposure may warrant further investigation for protectionof exposures during manual fire fighting procedures.

19

1200- uexposed1200

00-

L~LE

Q Boo E TiE (mEEm)EE

120

0:~~ exposede)

00EEEE E

L 40000

200

0020 30

TIME (min)

Fig. 11 b - Variable flow test-ericoal ulka average surface temperatures

12020

0C"4

LU 0

D Clu

LO

<8

ui(0183dl ONI70CL

< ........ ............ ........................................ ................ . (D

E o

0- 0

I. o

0 0 0 0 ( 0 0 0 0 D 0 D 00 00000000 000

(3)ain°joadw'a-L

21

LA-o

z- G

i~D

LL)

I--DD

L

U 0 LO 0 LO 0CNT-It

zw/m~i) Xfl iV3H iNVICOV8

22

7.2 Horizontal Boundary Cooling Tests

The results of the horizontal boundary cooling tests are listed in Table 1.Surface temperature measurements for these tests can be found in Appendix A. Theresults of this test series are similar to those recorded during the variable flow tests.Application rates greater than or equal to 2.04 Ipm/m2 (0.05 gpm/ft) provide roughlythe same amount of cooling independent of the type and orientation of the nozzle.The ability to cool a heated surface dropped off dramatically for application rates lessthan the critical value as shown in the variable flow tests and in Test 131. All testsconducted with application rates greater than or equal to 2.04 Ipm/m2 (0.05 gpm/ft2)reduced the unexposed surface temperature of the deck to 1 00C (21 2F) or below.Application rates above the critical value provided only a minimum difference in resultsas shown by Test 126. In Test 126, with the application rate nearly nineteen times thecritical value, the unexposed surface temperature was only 250C (450F) less than theaverage temperature produced using the critical application rate. These higherapplication rates produced a layer of water on the deck resulting in uniform coolingindependent of the system or technique used to apply the water. As a result of theresidual water buildup for application rates greater than the critical value, theapplication technique analysis became unnecessary. Application rates below 2.04Ipm/m2 (0.05 gpm/fte) were found to have only a minimum effect on compartment anddeck surface temperatures.

7.3 Vertical Boundary Cooling Tests

The results from the vertical boundary cooling tests are listed in Table 2.Surface temperature measurements for these tests are found in Appendix A. A fannozzle (FF 125145) demonstrated the highest application efficiency throughout thistest series. The spray pattern produced by this nozzle is best described as a thin, flatsheet. This nozzle was evaluated at three orientations. Maximum cooling wasachieved when the nozzle was oriented to provide complete bulkhead coverage andthe water droplets remained on the hot surface for a sufficient time to absorb therequired heat energy. A description of the orientations along with the correspondingresults are listed below.

Center Sideways. With this nozzle and orientation, the critical application ratedetermined in the variable flow and horizontal boundary cooling tests producedsimilar results. During these tests, the unexposed surface temperatures weredecreased to temperatures approaching 11 5oC (2390F) for all three tests. Thisnozzle and orientation produced the highest application efficiency for thesystems evaluated in this test series (Tests 104, 112, and 117).

Deck Up. When the nozzle was moved to the deck, the application efficiencydecreased. With the nozzle oriented in this fashion, the unexposed surfacetemperatures were measured to be 3000C (5720F) (Test 107). These results aredramatically higher than those recorded when the nozzle was mounted to theside. The decrease in cooling was the result of the water droplets falling down

23

0. c

~ Sn n Sn

Go 0 )

E Sn Ce-- q

E y

Ce) 0e 0 LO

0 0

I'0 00, 09

.9 a -cr - -.U)n

0 0) V) 0

S2E0

0

m --

N CD,0 0 9

o a) C6 '

a:00 )

Ir - 0) 0

N01

0 0. 0.

b--0 0h V

.0 n 0~ _

- >~ >Sn E El

0E N 0CM ~ 0 ~ c~0 u C D) c- a) co) 1-D C

0n (0

24

CL CL

V0 C#) C') CV')

an C75EW a V-0) 0 CO

m 0 - o

61 0 I) 0 61O in

>n in anO

Y CY -in C - t

<) o IV V 1 O 0 W' 1W

C') 0~ at)0

E a s 2.

0 I,- C CM coi )

No0 0 0

> at) at) U') In Of CY) CY C')<) 0w 09 -~ CM CY C~ Y C COj C

0 cc

CO N O 0 l) L

CL ,. aa a. 0

CL NE 0)i 0)i 0) 0 C') CO

.2<cr

4) (') 0 0 CO l . C

.2 ~I C) - N

I-C

N gn~ co coco C

t5 It) in 0) co 0) 0) v C

to I O tn cr) m' cr) m ) C M N

a) 0) M 0 c CCtWc tM M C CO CL (

LOLL~ j LI.. W CDc 0 0 M U D I 6 - U0 9n ~ .

>~ Oh D vl co vr 9 vo CL a-L

E 3': - 4 ~ *N _m V.- X:)U- La CL g.C. COO W) u a

c 0 -E M l jWMw-~ N 0 O

9- LL (D U-4 9M L - 0) LL9

25

CL E. - -

E 0~ __ N

In 0 U)

Lon C')

E i:1 -.0W 0-- 0

N NY

0 b o02 0 0 0 0

inC

0

U) a:) ) ~ iUc0 0 0 0 N 0

.0 0 65 C d c 6

m cm a' ) NM

-o a: -c-

15 CR ' ( '

t. CD

00

Nw

0) 0)

LL 0 m U)

c) CM C.) ln

0 I~ U. . 0 0 - o LL 0

26

the bulkhead, deflecting the pattern spraying upward. The net result wasincomplete coverage of the heated surface by the water spray pattern.

Ceiling Down. With the nozzle mounted at the ceiling, the unexposed surfacetemperatures were measured to be 3800C (716*F) (Test 106). Thisconfiguration proved to have the lowest application efficiency of the threeorientations. When the nozzle was mounted in the ceiling, the water dropletsremained on the surface the shortest amount of time producing the leastamount of cooling.

The nozzles that produced conical water spray patterns (TF1OFC, TF14FC)provided inadequate cooling (Tests 101, 102, 113) for low application rates butimproved substantially for higher application rates. Even at higher application rates,the conical water spray nozzles were unable to match the cooling produced by the fannozzles. Even at double the critical application rate, 3.66 Ipm/m 2 (0.09 gpmfte) (Test113), these nozzles were only able to reduce the unexposed bulkhead surfacetemperatures to 2000C (3921F).

The inefficient cooling produced by the fine atomizing nozzle P120 in Test 105suggests that all the water droplets were not impacting the hot bulkhead surface, butinstead, were being deflected by the thermal updraft of hot gases and steam. Even attwice the critical application rate (Test 110), the fine mist proved to be inadequate tocool the bulkhead (unexposed bulkhec1 temperatures of 3000C (5720F)). Thesurprisingly inefficient cooling produced by the fine droplets suggests that only thelarger drops have sufficient mass and resulting momentum to penetrate the thermalupdraft and strike the hot surface.

In summary, maximum application efficiency was obtained by applying watertangentially to the boundary surface in sheets or large droplets with nozzles thatproduce a fan shaped water spray pattern. Maximum cooling was achieved when thenozzle was oriented to provide complete coverage of the bulkhead by the water spraypattern, and the water droplets remained on the heated surface for a sufficient time toabsorb the heat. Full cone nozzles spraying perpendicular to the heated surfaceprovided some cooling for medium size droplets at higher application rates, while fineatomizing nozzles did very little to reduce the bulkhead surface temperature.

An interesting observation was made when comparing the bulkhead and decksurface temperatures of the lower application rates to the vari-nozzle baseline data.The lower application rates produced minimal amounts of visual steam and reducedonly the unexposed surface temperature, while the higher application rates producedsubstantial amounts of steam and reduced the temperature across the plate, bothexposed and unexposed surfaces (Fig. 14). The higher flows produced more steamas a result of additional energy being removed from the plate. This additional energyremoval and increased steam production are only beneficial in slowing recovery(reheat) times which was a factor in developing manual cooling techniques usingstandard Navy hardware. Reducing the back side surface temperature had no effecton the fire compartment temperature when the fire is still burning as shown in the

27

12001100--

01000900 0~Z

--- 800 FIRE SIDESURFACE TEMPS. 0o-700-

" 600 :

®. 500E 400-COOLING SIDE

300- SURFACE TEMPS.

200100

0 5 10 15 20

Time (min)

Fig. 14a - Fire compartment east wall average temperaturesat low application rate (3 gpm)

1200O

1100. o

01000 0

,_, 900 0.u FIRE SIDEv 800 SURFACE TEMPS.

0

700 0

Z 600CL 500-E 400

' 300

200 COOUNG SIDE

1 00 SURFACE TEMPS.

0 ,0 5 10 15 20

Time (min)

Fig. 14b - Fire compartment east wall average temperaturesat high application rate (60 gpm)

28

Total Periphery Cooling Test. However, if the fire was less severe or had already beenextinguished, reducing the exposed surface temperature would reduce the timerequired to restore the tenability of the fire compartment.

During the tests conducted with application rates of 2.04 Ipm/m2 (0.05 gpm/ft2),it was observed that when the nozzle was first activated, very little visual steam wasproduced. As the plate temperature was reduced, the steam production increased.None of the previously incorporated calculation procedures could explain thisphenomenon. A literature search on water droplet evaporation on hot surfacesproduced an explanation. Water vaporizes by three mechanisms; evaporation,nucleate boiling, and film boiling. Evaporation occurs when the water drop rests on asurface of less than 10(C (212*F). Nucleate boiling which is best described as auniform rolling boil, produces the highest vaporization rate and occurs when the platetemperature is between 1000C (2120F) and 2500C (4820F) depending on the platematerial. The region above 2500C (4820F) is known as the Uedenfrost Transition inwhich film boiling occurs. In the film boiling range, a thin vapor barrier formsunderneath the droplet lifting it off the surface which substantially reduces the heattransferred to the droplet.

Studies conducted by Michiyoshi and Makino [6] for the development ofnuclear reactors, illustrated the heat transfer characteristics of a droplet resting on aheated surface as a function of surface temperature and surface material. Figure 15shows the evaporation time of a single pure droplet of water setting on a heatedsurface as a function of surface temperature. The nucleate boiling regime is illustratedby the significantly faster evaporation times represented by the minimum values onFig. 15. Although it appears difficult to take advantage of this phenomenon, theinformation was incorporated in the development of techniques using standard Navyhardware.

7.4 Manual Cooling Techniques

A total of twelve tests were conducted in this test series. A summary of theresults is shown in Table 3. A wide range of application techniques and fire fighterpositions were evaluated during these tests. During tests with similar applicationtechniques, the higher flow nozzles produced lower bulkhead temperatures than thosenozzles having much lower flow rates. However, along with the lower temperaturescame an increase of water build-up in the compartment.

Incorporating the data from these tests on both cooling and reheat rates, a setof application techniques was developed for both the 360 Ipm (95 gpm) and 114 Ipm(30 gpm) vari-nozzles. These techniques were developed around the criteria ofkeeping the temperatures of the vertical bulkhead below 2500C (4820F) while applyingminimum amounts of water. The vertical bulkhead was selected due to both its strongdependence on application technique and consequently, techniques developed for thevertical bulkhead will work for horizontal decks as well. The selected temperaturerepresents a heat flux from the boundary of less than 5.G !'.m (0.45 BTU/ft2 sec)which would be inadequate to spontaneously ignite Class A materials (paper, wood,

29

IDo= 2.84 mm

UP Eg.(9),Baumeister et al/2/

1Eg.(1 2),o =6.2510 Doexp0.02870 -Two)

C0

0 0

0 04)

E 10 0o00-0 o copper

e brass* carbon steel

6 .* stainless steel

100 200 300 400 500

Surface temperature (C)

Fig. 15 - Vaporization rate as a function of platetemperature (Ref. 6)

30

-C0 N0 -N-C

o Nl N

CL- 0 A 0 0 0 0m 00 0 0 0 0 0

E - 0 CV) C'

Q~N 0 0 CO to ) 0 0a

C.)

E ~~ w. c (0c

) aS u, u 0 U, 0 0 uSC~N vO Iq

C~r- F3

- -- -- ,- .-

aw N.. v N - ') ') o C) -(

V~0 ) U) 0 0 cj co C

E -r C (0 O

a) a)c I C) N) PU) - P0- E 0 d C, u (0C f 7

n En m~ mw N IT E w E

0) 0M m 04 M M 0N M

U, 0 L 0 In 0 0

0 ~ ~ a 0) C' ' ') ') CD) CD)

CO cc- CC WOc

0 > > > >1

clothing) in dry air, much less when the combustibles are saturated with water. Thistemperature is also the lower limit for the Liedenfrost Transition region, where, for anyhigher temperatures, water itself begins to lose efficiency in cooling as stated earlier.

Temperature measurements recorded during the cooling of the boundariesusing the two developed techniques are shown in Fig. 16. These techniques aredescribed as follows.

The cooling technique developed for the 360 Ipm (95 gpm) vari-nozzle is statedas follows. While staying low or lying on the deck (if appropriate) a tolerable distancefrom the heated surface, the nozzleman should begin wetting the combustibles in thecompartment by sweeping with a 300 fog pattern. It is recommended that the firefighter position him or herself either in a doorway spraying into the heatedcompartment or behind a large piece of equipment to help protect the fire fighter fromdirect exposures to heat and steam. During the sweeping of the combustibles, shortsweeps of the heated surface should be conducted until the amount of steamproduced begins to make the space untenable for the fire fighter. In many cases, anatural inflow of cool air and outflow of steam will be established allowing thecompartment to remain tenable. If the position of the fire fighter is not jeopardized,the nozzleman should aggressively cool the boundary for approximately two and one-half minutes. This, in most cases, will reduce the boundary surface temperature tobelow 1500C (3020F). Short 15 second water bursts sweeping a 30" fog patternacross the bulkhead or deck every three minutes will ensure that the surfacetemperature never rises above 2500C (482"F).

The technique developed for the 114 Ipm (30 gpm) vari-nozzle is similar to theone developed for the 360 Ipm (95 gpm) vari-nozzle Initiating the cooling aspreviously described, the nozzleman should aggressively cool the boundary for fiveminutes using a 300 fog pattern. Short 15 second water bursts every two minutes willmaintain the boundary surface temperature below 2507C (482TF).

An analysis was also conducted on two low flow nozzles (a garden hose nozzleand a "homemade" nozzle made from a 1.9 cm (0.75 in.) pipe cap). Both nozzles hadK factors in the range of 0.5-1.0. The garden hose nozzle was selected for thisevaluation due to its availability onboard ships for cleaning purposes. The homemadenozzle was designed to simulate a low flow vari-nozzle. The homemade nozzle wasmade from a 1.9 cm (0.75 inch) threaded pipe cap with nine 0.13 cm (0.05 inch)diameter holes drilled into it. The holes were oriented to simulate the 30" fog patternof the vari-nozzle. The garden hose provided adequate cooling reducing the surfacetemperature to 180C (356"F) but suffered in reach characteristics. The garden hosenozzle only had a reach of 2-3 meters (6.7 feet) while producing a 300 water spraypattern. The homemade nozzle proved to have a greater reach, 5 to 6 meters (15-20ft), and had a much better cooling efficiency. The homemade nozzle cooled thebulkhead to 110"C (230F). Although these two nozzles provided adequate cooling atlow flow rates for this application, both of these nozzles would be ineffective againstsubstantially larger bulkheads or decks.

32

1200

1100-1000 o kA

900 0S 00 FIRE SIDE

700 SURFACE TEMPS.

0 600 in 0

500-E 400-

I- 300-

200 COOLING SIDE

100 SURFACE TEMPS.

0 ,

0 5 10 15 20 25

Time (min)

Fig. 16a - Fire compartment east wall average temperatures when coolingwith 30 gpm vari-nozzle (Test 121)

12001100..1000 0 2)

900 a)FIRE SIDE

1 800 SURFACE TEMPS.0S700-013

0 600 U 'J

S500--iE 400 I

300200 COOLING SIDEii ,

SURFACE TEMPS.

100

0 5 10 15 20 25Time (min)

Fig. 16b - Fire compartment east wall average temperatures when coolingwith 95 gpm vari-nozzle (Test 122)

33

8.0 SUMMARY AND CONCLUSIONS

1 . An application rate of 2.04 Ipm/m2 (0.05 gpm/ft) was verified to be minimum forcooling surface temperatures of both horizontal and vertical boundaries to100-C (212-F).

2. The temperature of the fire compartment was not reduced by aggressivelycooling (high application rates) of all fire compartment boundaries as long asthe fire remained burning.

3. A low flow water curtain (11.4 pm (3 gpm)) was found to reduce radiant heatby 75%.

4. At application rates of 2.04 Ipm/m 2 (0.05 gpm/ft) or above, cooling of horizontaldecks was independent of application technique.

5. Vertical boundary cooling was strongly dependent on application technique.Maximum cooling efficiency was achieved when the water was applied in sheetsor continuous sprays tangentially to the heated surface. The nozzle must alsobe oriented to provided complete surface coverage by the water spray and thewater must remain on the surface a sufficient time to absorb the requiredenergy. A nozzle producing a fan shaped water spray pattern mounted at theside of the bulkhead, one-half the vertical distance up the wall, spraying acrossthe heated surface produced superior results in this test series. Nozzlesproducing a conical shaped water spray pattern applied perpendicular to thesurface provided inadequate cooling for low application rates but improved withincreased application rates. The higher application rate was required due tothe water droplets (streams) bouncing off the heated surface. Fine atomizingnozzles applied perpendicular to the heated surface proved to be inadequatefor cooling of steel bulkheads. The fine drops lacked the needed momentum topenetrate the thermal updraft of steam and hot gases to adequately impact theheated surface.

6. Water was found to be more efficient at removing heat energy when wallsurface temperatures were between 1000- 2500C (2120- 4820F), i.e., in theLeidenfrost Transition region.

7. At low application rates, only the front surface of the boundary was cooled.Minimum visible steam was produced. At high application rates, both the frontand back surfaces of the boundary were cooled and significant amounts ofsteam were produced. The increased steam production was attributed to theability to remove the heat stored in the steel plate.

8. Techniques for effectively cooling boundaries to prevent fire spread weredeveloped for two standard Navy vari-nozzles and are listed as follows:

34

360 1pm (95 gpm) vari-nozzle - After the wetting of the combustibles inthe compartment, aggressively cool the boundary by sweeping with a300 fog pattern for two and one-half minutes or until conditions becomeuntenable for the fire fighter. Short 15 second water bursts every threeminutes should be more than adequate to keep the bulkhead cool andprevent fire spread.

114 Ipm (30 gpm) vari-nozzle - After the wetting of the combustibles inthe compartment, aggressively cool the boundary by sweeping with a30° fog pattern for five minutes or until conditions become untenable for

the fire fighter. Short 15 second water bursts every two minutes shouldbe sufficient to keep the bulkhead cool and prevent fire spread.

9.0 RECOMMENDATIONS

It is recommended that the results of these tests, as summarized in AppendixB, be included in the next revision of NSTM 555.

35

10.0 REFERENCES

1. J.T. Leonard, et al., "Post-Flashover Fires in Simulated ShipboardCompartments: Phase I-Small Scale Studies," NRL Memo Report (inpreparation).

2. Naval Sea Systems Command," Naval Ships Technical Manual, Chapter 555,Section V (proposed)," prepared by SEA 56Y52, 2 March 1989.

3. D. Gross and W. Davis, "Burning Characteristics of Combat Ship Compartmentsand Vertical Fire Spread," Center for Fire Research, National Institute ofStandards and Technology, NISTIR 88-3897, December 1988.

4. F. Sellars, 'Vital Space Boundary Fire Protection Concepts," letter to D. Kay,NAVSEA 56Y52, MPR 55-193, 13 September 1988.

5. D. Beene, "Evaluating Adjustable Pattern Nozzles for Engine Room FiresOnboard Coast Guard Cutters," US Coast Guard Marine Technical andHazardous Material Division Report CG-M-3-88, Groton, CT, June 1988.

6. I. Michiyoshi and K. Makino, "Heat Transfer Characteristics of Evaporation of aLiquid Droplet on Heated Surfaces," Int. J. Heat Mass Transfer Vol. 21, pp. 605-613, 3 June 1977.

36